Measuring linearity by area-length of convex hull
Lim Ai Hui, Kong Voon Pang · AIP conference proceedings · 2019
Effective linearity measures are useful in image processing. Sets of points with relatively high linear pattern are taken as support region. In this work, we introduce a linearity measure known as convex hull area-length ratio to obtain support region for boundary points of an image. The measure is invariant under transformations such as scaling, rotation and translation. The algorithm is using simple computation and resistant to protrusion in data set. Dominant point detection is the extension application after the support region determination. The capability of the measure is presented graphically with boundary points of images.